| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
* Medicity Research Laboratories, and Department of Medical Biochemistry and Molecular Biology, University of Turku, FIN-20520 Turku, Finland;
Turku Graduate School of Biomedical Sciences, University of Turku, FIN-20520 Turku, Finland;
Department of Pediatrics, Turku University Central Hospital, FIN-20520 Turku, Finland;
Department of Pathology, University of Turku, FIN-20520 Turku, Finland;
# Department of Medicine, Turku University Central Hospital, FIN-20520, Finland;
Department of Molecular Medicine, A. I. Virtanen Institute, University of Kuopio, FIN-70210 Kuopio, Finland;
Department of Medical Biochemistry, Ehime University School of Medicine, Shitsukawa, Shigenobu-cho, Onsen-gun, Ehime 791-0295, Japan; and
|| Molecular/Cancer Biology Laboratory, The Haartman Institute and Biomedicum Helsinki, University of Helsinki, FIN-00014 Helsinki, Finland;
** Department of Oncology, Turku University Central Hospital, FIN-20520, Finland
2Correspondence: Medicity Research Laboratories, University of Turku, Tykistökatu 6 A, FIN-20520 Turku, Finland. E-mail: klaus.elenius{at}utu.fi
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: angiogenesis • cancer • ErbB • HB-EGF • Herceptin
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
. If the support provided by the SMCs is inadequate or blocked, vessels become dilated and leaky or start to regress (2
, 4)
. These phenomena lead to poorly functional blood vessels that have an impaired capacity to deliver oxygenated blood or eliminate waste products. Thus, the EC–SMC interaction is one of the targets in the quest for strategies to promote or inhibit angiogenesis (3)
.
Soluble growth factors are important mediators of the paracrine interactions between ECs and SMCs. Well-characterized examples include platelet-derived growth factor-B (PDGF-B) and the angiopoietins (Ang). Null mice with disrupted genes for PDGF-B, or its receptor PDGFR-ß, die perinatally and indicate a lack of recruitment of PCs and vascular SMCs by ECs (5
, 6)
. Mice lacking Ang-1 or its receptor Tie2 die during midgestation with cardiovascular defects explained by a failure of the ECs of the blood vessels and endocardium to signal and associate with the adjacent vascular SMCs and cardiomyocytes, respectively (7
, 8)
. The phenotypes of Ang-1 -/- and Tie2 -/- mice resemble that of mice overexpressing another member of the Ang family, Ang-2, under an EC-specific promoter (9)
. Based on these in vivo findings and cell culture studies (9)
, it has been suggested that Ang-2 functions as a natural antagonist for Ang-1 effects on blood ECs, competing for binding to the same Tie2 receptor. Tie2 has been documented to be an EC-specific receptor not expressed by other cell types (10)
. This has led to the hypothesis that Ang-1 stimulates the expression or secretion of an EC-derived factor which then recruits SMCs by chemoattraction (2)
. The observations that the vascular defects in Ang-1 -/- or Tie2 -/- mice occur earlier in development than the vascular defects in PDGF-B -/- or PDGFR-ß -/- mice suggest that this Ang-1-induced SMC chemoattractant may be distinct from PDGF-B (5)
.
The ErbB receptors form a subfamily of receptor tyrosine kinases that consists of four members: ErbB1 (also known as EGF-receptor or HER1), ErbB2 (c-Neu, HER2), ErbB3 (HER3), and ErbB4 (HER4) (11
12
13
14
15)
. The ErbB receptors specifically interact with approximately a dozen epidermal growth factor (EGF) -like growth factors such as EGF, transforming growth factor-
(TGF-), amphiregulin (AR), heparin binding EGF-like growth factor (HB-EGF), betacellulin (BTC), epiregulin (EPR), epigen, and the neuregulins (NRG-1, NRG-2, NRG-3, and NRG-4) (16
17
18
19)
. Gene targeting experiments have demonstrated that the ErbB signaling system functions in a paracrine fashion in the communication between endocardial ECs and myocardial muscle cells in the heart. Homozygous mice with disrupted ErbB2, ErbB4, or NRG-1 genes all die around embryonic day 10 with defects in the formation of projections of the heart ventricles, called trabeculae (20
21
22)
. Since NRG-1 is expressed solely in the endothelial lining of the heart (endocardium), and ErbB2 and ErbB4 in heart muscle (myocardium), these findings indicate that a paracrine signaling system consisting of endocardium-derived NRG-1 signaling via ErbB2/ErbB4 heterodimer in cardiomyocytes is necessary for normal heart development.
Here we address whether ErbB receptors and their ligands also play a role in the paracrine interplay between blood vascular ECs and SMCs. Our results suggest that HB-EGF expressed by ECs can recruit SMCs by signaling via ErbB1 and ErbB2 in SMCs. Moreover, Ang-1 is capable of inducing the expression of HB-EGF in ECs as well as the potential of ECs to stimulate SMC migration, suggesting an indirect mechanism by which Ang-1 recruits SMCs. These findings provide new evidence about molecular signals that regulate EC–SMC interactions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
was produced in SF21 cells using BAC-TO-BAC baculovirus expression system (Gibco BRL, Grand Island, NY, USA) and was purified by two successive column chromatographies on heparin-Sepharose and C4 reversed phase HPLC.
To produce recombinant Ang-1 and Ang-2 fusion proteins, expression vectors encoding Ang-1 or Ang-2 coupled to the Fc fragment of human immunoglobulin gamma (Ang-1-Fc or Ang-2-Fc, respectively) (24)
were transfected to HEK293 cells using Lipofectamine reagent (Gibco BRL). Transfected and nontransfected HEK293 cells were grown to confluency, cell layers were washed with PBS, and the cultures were continued in serum-free DMEM for another 7 days. The conditioned media were centrifuged and the supernatants were concentrated 20-fold using Amicon Concentrator with YM-10 ultrafiltration membrane (Millipore, Bedford, MA, USA). Presence of Ang-1-Fc and Ang-2-Fc in the supernatant was confirmed by Western analysis using rabbit anti-human Fc antibody (Zymed, San Francisco, CA, USA) (see Fig. 6
A).
|
For adenoviral protein expression, HUVECs were grown to 70% confluency and exposed (400 MOI) to an adenovirus encoding angiopoietin-1 (AdAng-1) (25)
or ß-galactosidase (AdLacZ) (26)
for 2 h in RPMI containing 1% human AB serum (Finnish Red Cross). After a wash with PBS, cultures were maintained in RPMI containing 10% AB serum (for Western analysis) or in plain RPMI (to produce conditioned medium for migration assays).
Neutralizing antibodies for human HB-EGF (R&D) and ErbB2 (Herceptin; Roche, Nutley, NJ, USA) were used at final concentrations of 1 µg/mL and 10 µg/mL, respectively. CRM 197, a specific inhibitor of HB-EGF (27
, 28)
(Sigma, St. Louis, MO, USA), was used at the concentrations indicated in Fig. 5
. The ErbB1 inhibitor PD 153035 (Compound 32; Calbiochem, San Diego, CA, USA) was used at a final concentration of 1 µM.
|
Cell culture
Human umbilical vein endothelial cells (HUVEC) were prepared as described (29)
. HUVECs were cultured on dishes coated with 2% gelatin in RPMI medium supplemented with 10% AB serum, 1% glutamine/penicillin/streptomycin (GPS) supplement (Irvine Scientific, Santa Ana, CA, USA), 1.4 IU/mL heparin (Heparin LEO, Leo Pharma, Ballerup, Denmark), and 0.02 mg/mL bovine endothelial cell growth factor (ECGF; Roche). HUVECs were passaged maximally three times. Primary bovine capillary endothelial cells (BCE) (gift from Dr. Judah Folkman, Children’s Hospital, Boston, MA, USA) were cultured on gelatin-coated dishes in DMEM supplemented with 10% fetal calf serum (FCS) (Sigma or Bioclear), 1% GPS, and 3 ng/mL basic fibroblast growth factor (bFGF; PeproTech, Rocky Hill, NJ, USA). The expression of Tie2 in ECs was confirmed by Western analysis using a rabbit polyclonal anti-Tie2 antibody (C-20; Santa Cruz, Santa Cruz, CA, USA). Human aortic smooth muscle cells (HASMC) were obtained from American Type Culture Collection (ATCC No. CRL 1999) or purchased from PromoCell (Heidelberg, Germany). HASMCs were cultured in F12K (ATCC) medium including 10% FCS and supplements as recommended by ATCC. Bovine aortic smooth muscle cells (BASMCs) were established from tissue strips prepared from bovine thoracic aorta, as described (30)
. HEK293 cells (gift from Dr. Mika Scheinin, University of Turku, Finland) and BASMCs were cultured in DMEM supplemented with 10% FCS and 1% GPS.
Analysis of conditioned culture media and recombinant growth factors using Boyden chamber migration assays
To produce conditioned media for the migration assays, confluent cultures were washed with PBS and maintained for 24–48 h in plain DMEM or RPMI medium without any supplements. The media were collected, centrifuged, and diluted in DMEM or RPMI to obtain a concentration series. Recombinant HB-EGF and NRG-1 were diluted in DMEM before testing for chemoattractive potential.
The cells analyzed in the migration assays were starved overnight in serum-free medium, washed with PBS, trypsinized, and suspended in DMEM or RPMI to a final concentration of 500,000 cells/mL. Fifty microliter samples of the cell suspension were analyzed for migration in response to chemoattractants (conditioned medium, serum or recombinant proteins) using Boyden chamber apparatus, as described (31
, 32)
. The migration assays were carried out for 5 h in 37°C.
To determine the effect of specific inhibitors on SMC migration, inhibitory reagents were either added together with the chemoattractant to the lower Boyden chamber wells (CRM 197; anti-HB-EGF antibody) or together with the cells to the upper Boyden chamber wells (ErbB1 inhibitor; anti-ErbB2 antibody).
RT-PCR analysis of EGF-like ligands
Five micrograms of total RNA was extracted from confluent HUVEC cultures using the RNAzol B reagent (Tel-Test), and transcribed to cDNA with mouse myeloid leukemia virus reverse transcriptase (M-MLV RT; Promega, Madison, WI, USA), according to the manufacturer’s instructions. PCR amplification was carried out using the following sense and antisense primers, respectively: 5'-TGTCCCCTGTCCCACGAT-3' and 5'-AGCCTTGCTCTGTGCCCA-3' for amplification of a 511 bp fragment of EGF; 5'-TGCGGGACCATGAAGCT-3' and 5'-TCTCAGTGGGAATTAGTCA-3' for a 638 bp fragment of HB-EGF; 5'-AAAATGGTCCCCTCGGCT-3' and 5'-TCTGGGCTCTTCAGACCA-3' for a 496 bp fragment of TGF-; 5'-GCTCCCATCGCCGATGA-3' and 5'-TTTGATGGCGCCATTCAGA-3' for a 537 bp fragment of EPR; 5'-TGCGAAGGACCAATGAGAG-3' and 5'-GCATGTTACTGCTTCCAGG-3' for a 522 bp fragment of AR; 5'-TAGTGATCCTTCACTGTG-3' and 5'-TTAAGCAATATTTGTCTCTTC-3' for a 470 bp fragment of BTC; 5'-ATGAAAAGCCAGGAATCGG-3' and 5'-AGTATCTCGAGGGGTTTGA-3' for a 502 bp fragment of NRG-1; 5'-AGCCAGACGGGACAGGTG-3' and 5'-AGGAGAGCTGGTTGATGCC-3' for a 379 bp fragment of NRG-2; and 5'-CTACAATGAGCTGCGTGTGG-3' and 5'-TAGCTCTTCTC CAGGGAGGA-3' for a 450 bp fragment of human ß-actin. All PCR reactions were carried out in a total volume of 50.2 µL including 5 µL of template (10% v/v of RT reaction), 31 µL sterile water, 1 µL of specific 5' and 3' primers (35 pmol/µl; see above), 1 µL Dynazyme DNAII polymerase (2.0 U/mL) (Finnzymes, Espoo, Finland), 5 µL 10 x Dynazyme buffer (Finnzymes), and 1.2 µL dNTP mix (10 mM; Finnzymes). The samples were denatured at 94°C for 3 min and subsequently cycled 35 times through 1 min steps of annealing at 65°C, extension at 72°C, and denaturation at 94°C. PCR products were separated in 1% agarose gels.
Western blot analysis of ErbBs, HB-EGF, and tyrosine phosphorylated proteins
Protein expression was analyzed by Western blot as described (32)
. The synthesis of ErbBs by SMCs was analyzed using the following primary antibodies: anti-EGFR for ErbB1, C-18 for ErbB2, C-17 for ErbB3, and C-18 for ErbB4 (all from Santa Cruz). The synthesis of HB-EGF by ECs was analyzed using a monoclonal neutralizing antibody (R&D) recognizing mature HB-EGF as well as two polyclonal antibodies (C-18 and M-18; Santa Cruz) recognizing the cytoplasmic tail of precursor HB-EGF. Peroxidase-conjugated goat anti-rabbit IgG (Jackson Immunoresearch Laboratories, West Grove, PA, USA), goat anti-mouse IgG (Cappel, Cochranville, PA, USA), and rabbit anti-goat IgG (Chemicon International, El Segundo, CA, USA) were used as secondary antibodies.
For tyrosyl phosphorylation analysis, confluent SMC cultures were starved overnight in serum-free DMEM, stimulated with 50 ng/mL of growth factors for 10 min at 37°C, and lysed in lysis buffer containing 1% Triton X-100, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride (PMSF), 10 µg/mL aprotinin, 10 µg/mL leupeptin, 1 mM sodium orthovanadate, 10 mM sodium fluoride, and 10 mM sodium pyrophosphate. Aliquots of the lysates corresponding to 75 µg of total protein were separated in 6% SDS-PAGE gels, transferred to nitrocellulose membranes, and analyzed by Western blot with anti-phosphotyrosine antibody (4G10; Upstate Biotechnology Inc., Lake Placid, NY, USA) or phospho-specific anti-ErbB1 and anti-ErbB2 antibodies [phospho-EGF receptor (Tyr1068) and phospho-HER2/ErbB2 (Tyr1248), respectively; Cell Signaling Technology, Beverly, MA, USA]. Primary antibodies were detected by peroxidase-conjugated goat anti-mouse IgG (Cappel) or goat anti-rabbit IgG (Jackson Immunoresearch Laboratories), and ECL (Amersham, Amersham, UK).
Northern blot analysis of HB-EGF
Confluent HUVEC cultures were starved for 6 h in RPMI containing 2% AB serum, washed with PBS, and treated with the test samples diluted in serum-free DMEM. Total cellular RNA was isolated using RNAzol B (Tel-Test). Ten microgram aliquots of total RNA were fractionated in 0.9% formaldehyde-agarose gels and transferred to Zetaprobe membranes (Bio-Rad, Hercules, CA, USA). The membranes were hybridized with a 32P-dCTP-labeled (Rediprime II; Amersham) 665 bp ApaI-SalI fragment of the full-length human HB-EGF cDNA (33)
. To control loading of the RNA samples, the membranes were rehybridized with a 32P-dCTP-labeled (Nick Translation; Roche Diagnostics) human 28S rRNA cDNA probe (34)
. After exposure to X-ray films, the hybridization signals were quantitated using MCID image analyzer (MCID M5, Imaging Research, St. Catharine’s, Ontario, Canada).
Real-time RT-PCR analysis of ErbBs and HB-EGF
Quantitative real-time RT-PCR analysis (TaqMan) of ErbB mRNA expression in HASMCs was performed (T. Junttila et al., unpublished results). To analyze the stability of HB-EGF mRNA after angiopoietin treatment, test samples were applied to starved HUVEC cells as described above. Eight hours later RNA synthesis was blocked by addition of 3.3 µg/mL actinomysin D (Sigma). Total RNA was then extracted 0, 1, 2, 4, and 8 h after the application of actinomycin D, as described above. cDNA synthesis and TaqMan real-time PCR analysis were subsequently performed (T. Junttila et al., unpublished results), using the following primers and probes for HB-EGF and the reference gene ß-actin: 5'-TTATCCTCCAAGCCACAAGCA-3' (HB-EGF forward primer), 5'-AGCCCCTTGCCTTTCTTCTTT-3' (HB-EGF reverse primer), and 5'-TTCCCGTGCTCCTCCTTGTTTGGTGT-3' (HB-EGF probe); 5'-ATCTGGCACCACACCTTCTACAAT-3' (ß-actin forward primer), 5'-CCGTCACCGGAGTCCATCA-3' (ß-actin reverse primer), and 5'-TGACCCAGATCATGTTTGAGACCTTCAACAC-3' (ß-actin probe).
Immunohistochemistry
To localize HB-EGF expression in developing vascular structures in vivo, formalin fixed paraffin sections of an aborted 10-wk-old human fetus were stained with IgG fraction of polyclonal chicken immune serum targeted against the cytoplasmic domain of human HB-EGF (1:800 dilution; gift from Drs. Rosalyn Adam and Michael Freeman, Children's Hospital, Boston, MA, USA). For the control sections, primary antibody was replaced by chicken preimmune IgG. Blood vascular ECs were localized from adjacent sections with a mouse monoclonal anti-CD34 (Becton Dickinson, Franklin Lakes, NJ, USA; 1:20 dilution) and SMCs with a mouse monoclonal anti--smooth muscle actin (-SMA) (Sigma; 1:40000 dilution). Epitopes for these primary antibodies were visualized with immunoperoxidase technique using biotinylated secondary antibodies (biotinylated goat anti-chicken IgG from Vector Laboratories, Burlingame, CA, USA; biotinylated goat anti-mouse from DAKO, Carpinteria, CA, USA), Vectastain Elite ABC Kit (Vector Laboratories) or ChemMate Detection Kit (DAKO), and diaminobenzidine tetrahydrochloride (DAKO and Vector Laboratories) as a substrate.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A). The maximal stimulatory effect obtained by HUVEC conditioned medium was 
60% of the maximal effect stimulated by FCS. On the contrary, conditioned medium from HASMCs did not stimulate the migration of HUVECs, although HUVECs responded to human serum by migration (Fig. 1B
). Similar results were obtained using bovine cells. BCEs stimulated the migration of BASMC (Fig. 1C
), but not vice versa (Fig. 1D
), although both cell types were responsive to FCS (Fig. 1C, D
). The maximal migratory effect obtained by BCE conditioned medium was stronger than that observed by stimulation with FCS (1.4-fold). These results suggest that both human and bovine ECs secrete soluble factors that stimulate the migration of vascular SMCs, but that SMCs do not produce factors that regulate the migration of ECs.
|
ECs express HB-EGF
To investigate the possibility that EGF-like ligands are EC-derived factors that stimulate the migration of SMCs, the expression of several EGF-like growth factors was analyzed by RT-PCR in HUVECs (Fig. 2
A). PCR products of expected sizes were detected only when HB-EGF or NRG-1 expression was analyzed (Fig. 2A
, lanes 3 and 8, respectively). No specific products were observed for TGF-, EPR, AR, BTC, or NRG-2 (Fig. 2
, lanes 4–7 and 9). The 400 bp band observed when EGF-specific primers were used (Fig. 2
, lane 2) was a PCR artifact, as demonstrated by the unexpected size (expected size 511 bp) and sequencing of the PCR product (data not shown). RT-PCR analysis using ß-actin specific primers served as a positive control for RNA quality and generated a single band of the expected size (Fig. 2A
, lane 10). The expression of HB-EGF mRNA in HUVECs was confirmed by real-time quantitative RT-PCR. HUVECs expressed HB-EGF mRNA at levels corresponding to 36% of ß-actin mRNA expression in the same sample. For comparison, HASMCs expressed HB-EGF mRNA at levels corresponding to 0.15% of ß-actin mRNA expression.
|
The expression of HB-EGF protein in ECs was further examined by Western blot analysis of detergent-soluble EC fractions. Four different anti-HB-EGF antibodies recognized a duplet migrating at 46–48 kDa from HUVEC lysates (Fig. 2B
). Two of the antibodies had been raised against different epitopes within the extracellular domain of human HB-EGF (Fig. 2B
, lane 1; and data not shown), and two against cytoplasmic sequences corresponding to human (Fig. 2B
, lane 2) or mouse (Fig. 2B
, lane 3) precursor HB-EGF. The antibody against mouse HB-EGF also recognized a single 35 kDa HB-EGF band in BCE cells (Fig. 2B
, lane 4). Mature HB-EGF was secreted to the conditioned medium of HUVECs, as demonstrated by ELISA using the two extracellular domain antibodies (data not shown). These results indicate that primary ECs express at least one ErbB-ligand, HB-EGF, at both the mRNA and protein level.
SMCs express ErbB1 and ErbB2
To determine whether vascular SMCs synthesize receptors for EGF-like ligands, the expression of the four known ErbB receptors in SMCs was analyzed by Western blot. HASMCs and BASMCs both synthesized proteins that were recognized by anti-ErbB1 or anti-ErbB2 antibodies and that migrated at the expected molecular mass for ErbB receptors, 170–180 kDa (Fig. 3
A, B, lanes 4 and 5). In contrast, no specific bands were recognized with anti-ErbB3 or anti-ErbB4 antibodies (Fig. 3C, D
, lanes 4 and 5). Lysates from NIH 3T3 transfectants expressing either ErbB1, ErbB2, ErbB3, or ErbB4 (35)
from T47D cells known to express all four ErbB receptors (36)
and from 32D cells known to lack detectable ErbB protein expression (37)
served as positive and negative controls (Fig. 3
, lanes 1, 2, and 3, respectively). Consistent results were obtained when HASMC ErbB expression was quantitated using real-time RT-PCR: the ErbB mRNA expression relative to ß-actin mRNA expression was 3.9%, 1.2%, 0.02%, and 0.05% for ErbB1, ErbB2, ErbB3, and ErbB4, respectively. These results demonstrate that vascular SMCs express receptors for EGF-like growth factors, i.e., ErbB1 and ErbB2, but not ErbB3 or ErbB4.
|
HB-EGF stimulates phosphorylation of ErbB1 and ErbB2 in SMCs
The binding of EGF-like growth factors to ErbB receptors leads to activation of the intrinsic receptor tyrosine kinases. To investigate the responsiveness of vascular SMCs to EGF-like ligands, the presence of tyrosine phosphorylated proteins in SMCs was analyzed after 10 min stimulation with 50 ng/mL of EGF-like ligands by Western blot using phosphotyrosine-specific antibodies (Fig. 4
A). EGF, HB-EGF, and BTC induced tyrosine phosphorylation of proteins in HASMCs and BASMCs (Fig. 4A
, lanes 2–4 and 8–10). Consistent with these ligands activating ErbB receptors on the SMC surfaces, bands with the most prominent increase in phosphotyrosine content (Fig. 4A
) were of the size (170–180 kDa) of ErbB receptors expressed in these cells (Fig. 3
, lanes 4 and 5). In contrast, NRG-1 and NRG-2 did not stimulate tyrosine phosphorylation in SMCs (Fig. 4A
, lanes 5, 6, and 11) over the basal level observed in unstimulated cells (Fig. 4A
, lanes 1 and 7).
|
To assess ErbB phosphorylation more directly, HASMC lysates were analyzed by Western blot with phospho-specific anti-ErbB1 and anti-ErbB2 antibodies. As expected, HB-EGF induced phosphorylation of ErbB1 and ErbB2 (Fig. 4B
, upper panels, lanes 2 and 5), whereas NRG-1 had no effect (Fig. 4B
, upper panels, lanes 3 and 6). The effect of HB-EGF, but not of NRG-1, on ErbB1 was also observed in efficient down-regulation of ErbB1 protein in response to ligand stimulation (Fig. 4B
, lower panels, lanes 2 and 3, respectively).
These data indicate that ligands capable of binding to ErbB1 (EGF, HB-EGF, and BTC) can activate receptors on SMCs, whereas ligands that need ErbB3 or ErbB4 for signaling (NRG-1, and NRG-2) cannot. No ligand is known to bind ErbB2 directly, and ErbB2 has been suggested to signal in heterodimeric complexes with other ErbBs, including ErbB1 (16)
. Thus, the observed specificity of the ligand responsiveness of SMCs (Fig. 4)
was in accordance with the expression of ErbB1 and ErbB2 in SMCs (Fig. 3)
. Taken together, these findings suggest that EGF-like ligands such as HB-EGF that are capable of activating either ErbB1 homodimers or ErbB1/ErbB2 heterodimers can stimulate intracellular signaling in vascular SMCs.
HB-EGF, ErbB1 and ErbB2 are necessary for EC-stimulated migration of SMCs in vitro
Analysis of EGF-like ligands expressed by ECs (Fig. 2)
, the ErbB expression pattern in SMCs (Fig. 3)
, and the tyrosyl phosphorylation of SMCs in response to EGF-like ligands (Fig. 4)
indicated that HB-EGF is the only ErbB ligand expressed by ECs that can mediate paracrine signaling from ECs to SMCs. To determine whether EC-derived HB-EGF is necessary for EC-stimulated migration of SMCs, conditioned media from ECs were supplemented with increasing concentrations of CRM 197, a specific inhibitor of HB-EGF (27
, 28)
. The media were then analyzed for their ability to stimulate SMC migration using Boyden chamber assays. CRM 197 inhibited the EC-stimulated migration of both HASMCs and BASMCs in a dose-responsive manner (Fig. 5
A, B). However, significant quantitative differences were observed when different EC and SMC types were tested: A CRM 197 concentration of 0.1 µg/mL was enough to totally block the migration of HASMCs stimulated by HUVEC conditioned medium (Fig. 5A
). In contrast, the highest CRM 197 concentration tested, 100 µg/mL, was able to block 65% of BASMC migration stimulated by BCE conditioned medium (Fig. 5B
). CRM 197 did not block basal SMC migration or migration stimulated by 1% FCS (Fig. 5A, B
), demonstrating specificity of the effect.
The results obtained with CRM 197 were further confirmed using an antibody that neutralizes the bioactivity of human HB-EGF (R&D). The neutralizing antibody blocked 90% of HASMC migration stimulated by the medium (Fig. 5C
). Consistent with the expression (Fig. 3)
and activation (Fig. 4B
) of HB-EGF receptors ErbB1 and ErbB2 in HASMCs, a small molecular weight tyrosine kinase inhibitor of ErbB1 (PD 153035) and an inhibitory antibody recognizing human ErbB2 (Herceptin) also blocked the HUVEC medium-stimulated migration (Fig. 5C
). The inhibition was total when the ErbB1 and ErbB2 inhibitors were administered alone or in combination (Fig. 5C
). These results indicate that HB-EGF is necessary for efficient EC-stimulated migration of vascular SMCs in vitro and that the response requires both ErbB1 and ErbB2.
HB-EGF is sufficient for stimulating SMC migration in vitro
To address whether HB-EGF alone is sufficient to stimulate the migration of SMCs in vitro, recombinant human HB-EGF was tested as a chemoattractant in the Boyden chamber assay. Recombinant HB-EGF stimulated the migration of HASMCs and BASMCs in a dose-dependent manner, with the maximal effect at 20 ng/mL (Fig. 5D
and data not shown). As expected from ErbB expression and activation analyses (Figs. 3
and 4)
, 20 ng/mL recombinant NRG-1 did not affect HASMC migration (Fig. 5D
). The neutralizing anti-HB-EGF antibody (Fig. 5D
), the ErbB1 inhibitor PD 153035 (Fig. 5E
), and the neutralizing anti-ErbB2 antibody (Fig. 5E
) efficiently blocked HASMC migration stimulated by recombinant HB-EGF, indicating that HB-EGF signaling via ErbB1 and ErbB2 was mediating the effect. The neutralizing anti-HB-EGF antibody did not have an effect on the migration stimulated by 1% FCS (Fig. 5D
), suggesting that the recombinant HB-EGF preparation was free of contaminating stimulatory factors and that the effect of the antibody was specific (Fig. 5D
).
Angiopoietin-1 stimulates expression of HB-EGF in ECs
Angiopoietins have been shown to regulate the recruitment of PCs or SMCs toward ECs during angiogenesis (4)
. Since all known angiopoietins function through the EC-specific Tie2 receptor, it has been proposed that they exert their SMC recruiting activity indirectly by stimulating the production of SMC chemoattractants from ECs (2)
.
To analyze the effects of angiopoietins on endothelial HB-EGF mRNA expression, fusion proteins of Ang-1 or Ang-2 coupled to the Fc fragment of human immunoglobulin gamma (Ang-1-Fc or Ang-2-Fc, respectively) were produced in HEK293 cells. The presence of Ang-Fc proteins in 20-fold concentrated conditioned media from Ang-Fc-producing cells, but not in control medium from wild-type HEK293 cells, was confirmed by Western blot (Fig. 6
A). Serum-starved HUVEC cultures were then exposed to Ang-Fc or control medium or to 750 ng/mL of recombinant Ang-2 (from R&D), for 8 h and analyzed by Northern blot using an HB-EGF-specific probe. Ang-1-Fc up-regulated the expression of a single 2.5 kb HB-EGF transcript (Fig. 6B
, lane 2). In contrast, neither Ang-2-Fc (Fig. 6B
, lane 3) nor recombinant Ang-2 (Fig. 6B
, lane 5) had a stimulatory effect on HB-EGF mRNA levels when compared with control medium from HEK293 cells (Fig. 6B
, lane 1) or DMEM (Fig. 6B
, lane 4), respectively. Quantitation and time course analysis of the HB-EGF Northern data indicated that 1) Ang-1-Fc stimulated HB-EGF mRNA expression maximally by 6.4-fold, 2) the maximal effect was reached at 8 h, and 3) Ang-2-Fc and recombinant Ang-2 down-regulated rather than stimulated HB-EGF mRNA expression in HUVECs within the time range analyzed (Fig. 6C
). Similar data were observed when HB-EGF mRNA was quantitated using real-time RT-PCR (data not shown). Furthermore, real-time RT-PCR analysis of the stability of the HB-EGF mRNA in HUVECs demonstrated similar half-lives (
1 h) for HB-EGF mRNA after an 8 h stimulation with either Ang-1-Fc or control HEK293 medium (Fig. 6D
), suggesting that Ang-1 regulates HB-EGF expression at a transcriptional level. Consistent with its function as an Ang-1 antagonist for blood ECs, purified recombinant Ang-2 (R&D) was able to block the effect of Ang-1-Fc medium on HB-EGF expression in a dose-responsive manner (50% inhibition at 100 ng/mL, 80% inhibition at 1000 ng/mL), indicating that the activity of Ang-1-Fc medium was specific.
To analyze the effect of Ang-1 on endothelial HB-EGF protein expression, 70% confluent HUVEC cultures were infected with adenoviruses encoding human Ang-1 (AdAng-1) or ß-galactosidase (AdLacZ). Three days after infection, the expression of HB-EGF protein was detected by Western blot. AdAng-1 up-regulated HB-EGF protein expression (Fig. 7
A, lanes 1 and 2) by 5.6-fold as demonstrated by densitometric analysis of the Western film (Fig. 7A
, columns). These data demonstrate that Ang-1 is capable of up-regulating the expression of HB-EGF in ECs.
|
Angiopoietin-1 enhances EC-stimulated migration of SMCs in a mechanism involving HB-EGF
To more directly assess the role of HB-EGF in mediating the effect of Ang-1 on SMC migration, serum-free conditioned media from HUVECs infected with either AdAng-1 or AdLacZ were tested for chemoattractive potential for HASMCs in Boyden chamber assays. AdAng-1 infection up-regulated the capacity of HUVECs to attract HASMCs by 3.2- to 5.6-fold in three independent experiments (Fig. 7B
). Similar observations were made when HUVEC conditioned media were analyzed after a 24 h treatment with 200 ng/mL recombinant Ang-1 (R&D; data not shown). When 1 µg/mL of the HB-EGF inhibitor CRM 197 was added to the conditioned medium from AdAng-1-infected HUVECs, the capacity of the medium to stimulate SMC migration was reduced by 60% (Fig. 7B
). These findings demonstrate that Ang-1 enhances SMC migration stimulated by ECs in vitro and suggest that a major proportion of this enhancement is mediated via EC-derived HB-EGF.
HB-EGF is expressed in ECs associated with vascular SMCs or PCs in vivo
To study whether HB-EGF is expressed by ECs in angiogenic tissues in vivo, paraffin sections of a 10-wk-old aborted human fetus were immunostained with a chicken polyclonal antibody against the cytoplasmic domain of the mature transmembrane form of human HB-EGF. A positive signal was observed in ECs in a majority of developing vascular structures, including arterial and venous vessels as well as capillaries, in several tissues (Fig. 8
A, E). As expected based on several previous reports, muscle tissues, including vascular smooth muscle, were slightly positive for HB-EGF. Identical staining patterns were seen when 1) adjacent sections were stained using an independent anti-HB-EGF antibody recognizing human HB-EGF (C-18) and 2) sections of mouse embryos (embryonic day 14.5) were stained with an antibody recognizing mouse HB-EGF (M-18) (data not shown). No specific staining was observed when chicken preimmune serum IgG fraction was used (Fig. 8B, F, J
) or when primary antibody was omitted from the protocol (data not shown).
|
To confirm the identity of HB-EGF-positive cells, ECs and SMCs/PCs were visualized from adjacent sections by staining with antibodies raised against CD34 (Fig. 8C, G, K
) and -SMA (Fig. 8D, H, L
), respectively. Strong HB-EGF immunoreactivity seemed to colocalize with ECs of vascular structures containing -SMA-positive SMCs or PCs. For example, ECs in capillaries and venules in the developing skin were HB-EGF positive (Fig. 8E
) and associated with scattered SMCs/PCs (Fig. 8H
). In contrast, similar-sized hyaloid vessels in the developing eye were significantly less stained, if stained at all, with anti-HB-EGF antibody (Fig. 8I
) and were not associated with any detectable -SMA reactivity (Fig. 8L
). This is consistent with the later regression of the hyaloid vessels from the developing vitreous (38)
and with the role of EC-derived HB-EGF in SMC/PC recruitment during angiogenesis in vivo.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 4) recombinant HB-EGF stimulated tyrosine phosphorylation of ErbB1 and ErbB2 in HASMC, and 5) reagents that specifically neutralize the activity of either HB-EGF, ErbB1, or ErbB2 blocked the SMC migration-inducing activity secreted by ECs. In addition, 6) Ang-1, an EC-specific growth factor necessary for efficient SMC recruitment in vivo (8)
, up-regulated the expression of HB-EGF mRNA and protein in ECs and 7) enhanced the SMC migration-inducing activity of ECs by a mechanism largely dependent on HB-EGF. Finally, 8) HB-EGF was expressed in ECs associated with vascular SMCs/PCs during angiogenesis in vivo. These data support a role for HB-EGF, and its receptors ErbB1 and ErbB2, in the recruitment of SMCs by ECs.
SMC recruitment to newly formed vessels has been proposed to involve soluble EC-derived chemoattractants that interact with specific receptors on SMC surfaces (2)
. We have analyzed the communication between ECs and SMCs by measuring the potential of either cell type to stimulate the migration of the other in vitro. It was demonstrated using Boyden chamber migration assays that conditioned media from primary ECs stimulated the migration of vascular SMCs but that conditioned media from SMCs did not stimulate the migration of ECs. These findings indicate that our in vitro experiments provided a relevant model to characterize the paracrine factors that ECs use to recruit SMCs.
It was demonstrated by RT-PCR that at least two EGF-like ligands are expressed by primary HUVECs: NRG-1 and HB-EGF. Of the two factors detected in HUVECs, NRG-1 is unlikely to mediate signaling from ECs to SMCs since 1) neither HASMCs nor BASMCs expressed the NRG-1 receptors ErbB3 or ErbB4. SMCs did not respond to recombinant NRG-1 by 2) protein tyrosine phosphorylation, 3) specific phosphorylation of ErbB1 or ErbB2, or 4) migration. Together with observations of the in vivo angiogenic effect of NRG-1 (40)
(E. Iivanainen et al., unpublished results), these results imply a role for NRG-1 in vascular biology that is distinct from stimulating recruitment of SMCs. Several lines of evidence, however, support a role of EC-derived HB-EGF in EC–SMC interaction. Both HB-EGF mRNA and protein have been shown to be expressed by HUVECs in vitro (41
, 42)
. Here we demonstrated that HB-EGF is expressed in ECs associated with SMCs/PCs during embryogenesis in vivo. HASMCs and BASMCs were also shown to express HB-EGF receptors and to respond to recombinant HB-EGF by protein tyrosine phosphorylation, migration, and proliferation (data not shown). These findings are in accordance with HB-EGF’s well-established role as a potent mitogen and chemoattractant for different types of SMCs (43)
, including vascular SMCs (33
, 44)
.
The immunohistochemical analysis indicated that some HB-EGF was expressed by vascular SMCs, and HB-EGF released from SMCs has been suggested to function as an autocrine SMC effector (45)
. However, there seems to be less HB-EGF made by SMCs than by ECs. Immunohistochemical analysis indicated more HB-EGF expression in the EC layer than the SMC layer in developing arterial structures; HB-EGF protein was barely detectable by Western blot under conditions that clearly gave a band for HUVEC-derived HB-EGF (Fig. 2B
; data not shown) and HASMCs expressed only 0.4% of the HB-EGF mRNA amount expressed by HUVECs in real-time RT-PCR analysis. Thus, ECs are expected to be a more important and quantitatively greater biological source for HB-EGF in regulation of directional SMC chemotaxis than SMCs themselves.
In spite of indirect indications (2
, 42
, 46)
, this report is to our knowledge the first in which the significance of HB-EGF and the ErbB signaling system in mediating SMC migration has been experimentally addressed. To analyze this, we tested inhibitors of the HB-EGF signaling pathway for their effects on EC-stimulated SMC migration. A specific HB-EGF inhibitor (CRM 197), a neutralizing anti-HB-EGF antibody, a small molecular weight inhibitor of ErbB1 (PD 153035), and an inhibitory antibody against ErbB2 (Herceptin) all blocked HUVEC medium-stimulated migratory response of HASMCs by 90 to 100%. These findings suggest that inhibiting HB-EGF-stimulated ErbB signaling could also be a way to interfere with EC–SMC interaction during pathological angiogenesis in vivo. Recent reports indicate that some of the clinical anti-tumor effects of the ErbB1 and ErbB2 inhibitors Iressa and Herceptin, respectively, could be explained by anti-angiogenic effects on tumor vasculature (47
, 48)
.
In contrast to HUVEC-stimulated migration of HASMCs, CRM 197 was significantly less potent in blocking BCE-stimulated migration of BASMCs. This was not an obvious outcome of bovine cells being less sensitive to CRM 197, since the ErbB1 inhibitor PD 153035 was also significantly less potent in inhibiting BCE-stimulated migration of BASMCs compared with HUVEC-stimulated migration of HASMCs (data not shown). Bovine cells are known to be sensitive to the toxic variant of CRM 197, diphtheria toxin (E. Mekada, personal communication), indicating that bovine HB-EGF indeed interacts with CRM 197. One could speculate that since BCE-conditioned medium efficiently recruited SMCs whereas BCEs seemed to express less HB-EGF protein than HUVECs (based on analysis with an antibody against mouse HB-EGF), factors other than HB-EGF might play a more important role in BCE-stimulated SMC recruitment compared with HUVECs. Taken together, these findings strongly support a central role for HB-EGF in mediating EC-stimulated recruitment of SMCs, but also indicate that the relative importance of HB-EGF in this signaling may vary between the heterogenic vascular beds. Elucidation of the relative contributions of HB-EGF and other factors, such as of PDGF-B, in the recruitment of SMCs in different vascular structures needs further in vitro and in vivo experiments in which various vessel types and ligands are analyzed simultaneously. For example, no information is currently available about the vascular phenotype of targeted HB-EGF -/- mouse (49)
.
Mice lacking Ang-1 or Tie2 genes or overexpressing the antagonist Ang-2 have defects in the developing vessels associated with a lack of recruitment of PCs and SMCs (7
8
9)
. Since angiopoietins are supposed to signal solely via the EC-specific Tie2 (4)
, these findings indicate an indirect mechanism in which Ang-1/Tie2 signaling stimulates the production of a soluble SMC chemoattractant, such as PDGF-B, from ECs (2)
. Mice lacking Ang-1 or Tie2 already have defects in the SMC coverage of large vessels around embryonic day 10, when these structures in PDGF-B- or PDGFR-ß-deficient mice still develop normally (5)
. These reports support the presence of EC-derived SMC chemoattractants other than PDGF-B and suggest that the expression of such a factor could be regulated by angiopoietins. Here we demonstrate that Ang-1 stimulated the expression of HB-EGF in HUVECs. In addition, conditioned medium from AdAng-1 or recombinant Ang-1-treated HUVECs stimulated more HASMC migration than control media, and HB-EGF neutralization significantly blocked the enhanced migration. These findings suggest that endothelial HB-EGF expression is under the regulation of a central effector of vascular maturation, Ang-1, and support a crucial role for HB-EGF in a pathway mediating SMC recruitment by Ang-1. One possible source for Ang-1 in vivo is mesenchymal cells surrounding developing blood vessels (50)
. This is consistent with the observation that HASMCs used in our experiments expressed considerable amounts of Ang-1 mRNA but significantly less Ang-2 mRNA (2.4% and 0.07% of ß-actin mRNA, respectively) when analyzed by real-time RT-PCR and that serum-free conditioned medium from HASMC up-regulated HB-EGF expression in HUVECs (unpublished observations).
The Ang-1/Tie2 signaling system has been suggested to interact with the EGF/ErbB system during the development of the heart. The heart phenotypes of mice deficient in Ang-1, Tie2, NRG-1, ErbB2, and ErbB4 are almost identical, each demonstrating a lack of myocardial trabeculae (7
, 8
, 20
21
22)
. Ang-1 secreted from the myocardial cells and signaling via the Tie2 in endocardial cells has been proposed to reciprocally communicate with NRG-1 expressed in endocardium and signaling via the ErbB2/ErbB4 heterodimer in myocardial cells. Taken together, these observations and the data presented in this paper support a model in which cross-talk between the ErbB and Tie receptor tyrosine kinase subfamilies plays a role in the development and maintenance of several different cardiovascular structures. Tissue specificity in this cross-talk may be achieved by differential expression of the members of the ErbB system, e.g., by NRG-1 signaling via ErbB2/ErbB4 in the heart and by HB-EGF signaling via ErbB1/ErbB2 in blood vascular structures.
Collectively, the results suggest that HB-EGF is the major EGF-like ligand expressed by ECs that participates in the paracrine signaling from ECs to SMCs. The receptors that mediate HB-EGF signaling in SMCs are ErbB1 and ErbB2, indicating that this signaling pathway can be inhibited by available ErbB inhibitor drugs. Furthermore, the finding that Ang-1 up-regulates the capacity of ECs to recruit SMCs in an HB-EGF-dependent manner suggests a novel indirect mechanism by which angiopoietins regulate the maturation of blood vessels.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 23, 2002. Accepted for publication April 22, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
